Recovering Positions Of Identical Matches From Multiple/ Pairwise Sequence Alignment

Question

[VS]

I would like to do a pairwise/multiple sequence alignment for a gene from two/three species and then record in a separate file the exact positions of perfect identity between the sequences. Is there a tool that does this already? Or is there any option in blast/clustal that could help in recovering this information ?

Will be very thankful for your time and responses.

[arnstrm]

Have you tried GBlocks? http://molevol.cmima.csic.es/castresana/Gblocks.html I know it can extract highly conserved regions from an alignment, but not sure if you can configure it to extract only perfect matches. Edit: here is the webserver: http://www.vardb.org/vardb/analysis/gblocks.html

Answer 1

[Alex Reynolds]

BLAT can return exact matches, by filtering for results where the query and target sizes (sequence lengths) match. Perhaps you could do an EMBOSS Needle alignment with a very high gap penalty (to help push down gapped alignment results) and then use BLAT to do the sequence position lookup on aligned results.

[VS]

Thanks for the answer! Any particular reason you recommend EMBOSS Needle over blast with similarly high gap penalty?

[Alex Reynolds]

I think a local alignment (e.g., BLAST) won't always give you the edges of the query sequence in a search match, but global (e.g., Needleman-Wunsch) will. Because you want an exact match, you probably want those edges to match, too.